End users of Web task, legacy applications, Apple (Apple is a registered trade mark of Apple, Inc.) or Windows (Windows is a registered trade mark of Microsoft Corporation) based operating systems create few automated processes, that is a series of steps that a user desires to execute on a computer (often at pre-defined times or by touch of a screen or with a click of a button or a mouse). These software processes could be of any type and for any purpose, e.g., business, IT, or requirements to validate software behavior, etc. In some instances, the user desires to create a task, assign properties to the task and to run the task on a local or a remote computer. Desirable applications may be the creation of adapters that enable application automation by collecting automation information; the locating of application controls and tracking changes between an older and newer version of an application; presenting the changes using an exception management model to a user, so that the user can provide feedback in a visual, instead of programmatic manner; storing and incorporating the changes so as to make the adapters resilient to application changes and upgrades.
Traditional automation techniques are limited in their ability to perform automation tasks precisely and reliably, since these techniques do not have prior knowledge of pre-existing system and application conditions. Additionally, the current processing load is often never considered in determining system response, which can be critical in reliably automating tasks on that system. We propose a system and method to enable reliable and fault-tolerant application automation by utilizing artificial intelligence and knowledge based techniques.